1094. A comparison of prioritization methods in pairwise comparisons with respect to a priority vector quality
Invited abstract in session TC-10: Pairwise comparisons and preference relations 1, stream Multiple Criteria Decision Aiding.
Tuesday, 12:30-14:00Room: Clarendon SR 1.06
Authors (first author is the speaker)
| 1. | Jiri Mazurek
|
| School of Business Administration in Karvina |
Abstract
Pairwise comparisons (PCs) constitute a fundamental part of many multiple-criteria decision-making methods designed to solve complex real-world problems. The methods’ input has a form of experts’ judgments (pairwise comparisons of alternatives, criteria, etc.) usually arranged into a PC matrix. The methods’ output is a priority vector – the vector of weights of all compared objects. When pairwise comparisons are inconsistent, different methods yield different results (vectors), and these results do not always satisfy a set of natural properties such as coherency, consistency, intensity, or efficiency. The aim of this paper is to compare selected prioritization methods with respect to the satisfaction of the desirable properties by numerical examples and Monte Carlo simulations for different matrix sizes, different levels of inconsistency of an input PC matrix, and also with respect to an incomplete information (missing PCs). The examined methods include, among others, the AHP (the eigenvector method), geometric mean method (GMM), simple tree method, Best-Worst method, Weighted Least Squares Method, Simple Column Sum Method Rank, or AHP-Express. Our results indicate that the AHP and GMM (without missing matrix elements) perform the best.
Keywords
- Analytic Hierarchy Process
- Decision Theory
- Decision Analysis
Status: accepted
Back to the list of papers